Task-dependent inflation of rapid motion priming effects due to task and response constraints.

Visual motion priming (VMP) arises when an unambiguous stimulus clearly moving in one direction (e.g., leftwards) induces the perception that a subsequent ambiguous motion stimulus is moving in the same direction. VMP can be observed with very brief stimulus presentations and inter-stimulus intervals (ISIs) of only tens of milliseconds. Anecdotal evidence suggests that, under such conditions, observers may be unable to reliably individuate the prime and ambiguous stimuli. We argue that, for VMP measurement tasks which rely on individuated perception of the motion direction of the two stimuli (e.g., same-different comparisons), this limitation may inflate estimates of VMP. Across two experiments, we observed priming effects with such tasks. However, at the shortest presentation durations and ISIs, observers were at chance for discriminating the motion direction of the initial priming stimulus and frequently reported perceiving only one stimulus. We argue that, under these conditions, participants may rely on task demands rather than distinct percepts to make responses, biasing them to report the same motion direction for both stimuli which is interpreted as VMP. Consistent with this account, analysis of data from alternative tasks that do not require explicit discrimination of the two stimuli still showed VMP, but with substantially reduced magnitude. These findings suggest that VMP magnitude estimates can be inflated by response procedures that depend on conscious discrimination of briefly presented motion stimuli separated by very short intervals. We therefore recommend using paradigms that minimise such task-related biases when estimating VMP under these conditions.